Several families of proteins are under investigation for the determination of the relationships between structure and function. Most of these protein families are eukaryotic nuclear proteins and have the property of binding to DNA or RNA, for example, the histone proteins, the HMG-1 box proteins, heat shock factors, ets proteins, HMG-I(Y) proteins, and ribosomal proteins. As an example, the HMG-1 box family of proteins is not highly conserved and is represented by over 100 examples in the sequence databases. We have identified a signature for this family of proteins and have classified them into several groups according to their sequence and functional relation. In the past few years, numerous proteins have been identified as containing a stretch of about 75 amino acids which are homologous to an abundant non-histone chromosomal protein HMG-1. These proteins bind DNA and bends it on binding or bind preferentially to bent DNA. Several of these proteins have been implicated in numerous nuclear functions including transcription, replication, and chromatin structure as well as transcription regulation in mitochondria resulting, in some cases, in such phenotypes as sex and mating type determination. For example, we have compiled and are maintaining a database of the HMG-1 box family of proteins and are analyzing the sequences to determine the phylogeny between these functionally widely-diverse proteins. The 3D structure of several HMG-1 box domains have been determined by multi-dimensional NMR. We have used one of these structure to model the other members of the family by a threading method. We have successfully produced models for the complete family of HMG-1 box domains. The conclusion drawn from these calculations is that these proteins are mostly likely to fold into similar conformations. An iterative motif search algorithm is being used to detect new and as yet unidentified motifs of several families of DNA-binding and RNA-binding proteins. We have identified several proteins which contain domains which could be related to the histone fold of nucleosome octamer histones. We are extending this study to include the ability to identify partners in a dimerization interaction. In another study, the relationship between the heat shock factor and ets families of proteins has been identified by the same technique. Similarly, other non-histone chromosomal proteins have been investigated, for example, many new unidentified open reading frames have been suggested to be either histone acetyltransferases or histone deacetylases by similar methods.